Levulinic Acid Hydrogenation Reaction using a Polymeric Catalytic Membrane Reactor By Michelle C. Soto Hernández NSF REU Program in Sustainable Bioenergy 2015 Department of Chemical Engineering Kansas State University Final Research Report This material is based upon work supported by National Science Foundation Grant: REU Site: Summer Academy in Sustainable Bioenergy; NSF Award No.: SMA-1359082, awarded to Kansas State University.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Membrane Reactors and Three-Phase Hydrogenation P a g e | 1
Levulinic Acid Hydrogenation Reaction using a
Polymeric Catalytic Membrane Reactor
By
Michelle C. Soto Hernández
NSF REU Program in Sustainable Bioenergy 2015
Department of Chemical Engineering
Kansas State University
Final Research Report
This material is based upon work supported by National Science Foundation Grant: REU Site: Summer Academy
in Sustainable Bioenergy; NSF Award No.: SMA-1359082, awarded to Kansas State University.
Membrane Reactors and Three-Phase Hydrogenation P a g e | 2
Abstract
The environmental degradation caused by the consumption of fossil and the increasing energy
demand is of great concern when taking in account social welfare. It is forecast that fossil fuel will not
be able to defray society’s demands. Efforts are being placed in the study of lignocellulose-biomass
conversion, which represents a possible solution to this problem since platform molecules can be
obtained from it. These type of molecules can be used to produce fuels and fuels additive.
Hydrogenation reactions are industrially relevant in petrochemical and petroleum refining and in
emerging biorefining. The reaction that was studied was the hydrogenation of levulinic acid (LA) to
produce gamma-valerolactone (GVL), a chemical platform. Previous studies had conducted this reaction
in a packed bed reactor. One of the problems presented in this kind of reactor is the mass transfer
limitation, hydrogen delivery to the catalyst sites. Membrane reactors can lessen the mass transfer
limitations by supplying hydrogen to the catalytic sites on the membrane surface. The main objective of
this project was to develop a polymeric membrane for the hydrogenation of LA.
Polytetrafluoroethylene (PTFE) was used as the support of the polymer Matrimid. Among the different
types of membranes developed, we focus on the reaction conducted with a membrane that consisted of
0.25wt% and 1.0wt% Matrimid dissolved in methylene chloride (DCM) and spin-coated onto the PTFE
membrane. Also, different types of solvents for the LA hydrogenation were evaluated such as glycerol,
water and gamma-butyrolactone (GBL). At constant temperature of 70oC, glycerol did not show any
yield of GVL after13-15 hours of reaction. When GBL was used as solvent for LA, a bronze colored
liquid was obtained. Since this chemical has similar properties to GBL, it was not able to be analyzed by
simple means. Water showed to be the most efficient solvent to produce GVL. Also, the PTFE
permeability for water and GBL was evaluated. GBL has a lower vapor pressure in comparison to water,
therefore allowing the reaction to be conducted at higher temperature without extreme permeability of
the liquid phase, as expected due to its high boiling point. In the case of water, 70C was chosen as
reaction temperature for further experiments, because produced a manageable water permeance through
the PTFE membrane.
The hydrogenation of LA was conducted supplying hydrogen from the lower side of the
membrane and from the upper side (in a continuous loop flow). Results have shown that the most
Membrane Reactors and Three-Phase Hydrogenation P a g e | 5
1. Background
1.1 Research Justification
The fast growing population, rapid modernization and industrialization have increased the
demand in energy and this demand is mostly met from a non-renewable resource such as fossil fuel. The
dependence on fossil fuel has serious implications in the economy and in the environment. We are
entering an era of diminish availability of petrochemical resources used to produce energy and chemical
materials needed by society [1].
The growing environmental degradation is observed locally, nationwide and globally, and has
attracted the attention and concern in the scientific environment. Policies for reducing energy use may
not ensure the desired level environmental quality along with the desired level of economic growth and
social welfare. The energy sector is dominated by the direct combustion of fuels, a process leading to
large emissions of CO2, making carbon dioxide (CO2) one of the most foremost greenhouse gases in the
atmosphere (Figure 1). CO2 produced by the burning of fossil fuel is of great concern in view of its
impact on global warming, which
concentration in the atmosphere is
increasing at a rate of 0.4% per year [4,
8]. Seeking improvement and developing
an alternative sustainable method of
producing energy is imperative. Abundant
biomass resource are a promising
alternative for the sustainable supply of
energy. Biomass is a renewable carbon
source that can be processed in an
integrated bio-refinery, in a manner
similar to petroleum in conventional refineries, to produce fuels and chemicals [2].
1.2 Lignocellulose to produce levulinic acid and GVL
Lignocellulosic biomass is an alternative renewable source of carbon that has potential to be
converted to a variety of fuels (Figure 2). Two of three components are cellulose which accounts for
Figure 1. Shares of global anthropogenic greenhouse
gas emission [4].
Membrane Reactors and Three-Phase Hydrogenation P a g e | 6
30-50 wt% of lignocellulose biomass and hemicellulose with 15-30 wt%. Cellulose can be converted to
glucose that can be used to produce platform chemicals, such as 5-hydroxumethyl furfural (HMF). HMF
can then be converted to levulinic acid (LA) and then to gamma-valerolactone (GVL). GVL is a
promising renewable platform molecule that can help replace petroleum derived chemicals [10].
HMF can then be converted to LA and then to GVL. GVL has excellent properties as solvent, is a
precursor of high value chemicals, fuels, is
consider a green solvent, plus it’s not
decomposed nor degraded with time and has
high boiling point. A green solvent is said to
be a substance that reduces environmental
hazard and posed less health risks. One of the
criteria that one must take in account when
producing a solvent by means of a chemical
reaction are its intermediates products. For
GVL, starting from cellulose, some of its
intermediate products are glucose, HMF, LA, and
formic acid, which are miscible in water, which facilitate its biodegradability. In addition, GVL can be
used as solvent to produce HMF, LA, and GVL from fructose. One possible pathway to produce GVL
are hydrogenation or dehydration of LA. In this case, we focus on the process of hydrogenation of LA
(figure 3) that produces an intermediate product call gamma-hydroxyvaleric acid, which ring-closes by
intermolecular esterification and loses a H2O molecule to produce GVL. [10]
Figure 3. Levulinic acid transformation to GVL.
Figure 2. Roadmap for the conversion of lignocellulose [11].
4-hydroxypentanoic acid
Membrane Reactors and Three-Phase Hydrogenation P a g e | 7
In order for this reaction to proceed, a catalyst must be used. It’s important to take in account the
nature of the solvent, and the support of the catalyst. Many catalyst have been used to produce GVL, but
the one that achieve the highest yield was Ru-based catalyst [10, 11]. As mentioned before, considering
GVL from the sustainable perspective, GVL is a chemical of great value. This chemical can be used in a
similar manner as ethanol, and can be hydrogenated to produce fuel additives (Figure 4) such as MTHF
with a 93% of selectivity at 98% GVL conversion using Cu catalyst, or 1,4 pentanodiol.
1.3 Utilizing membranes as phase contactors for three-phase hydrogenation
It’s imperative that every chemical reaction is carried out with high selectivity and high
Figure 4. Reaction pathways for the conversion of GVL into fuels, fuel additive and chemicals [10]. Figure 4. Reaction pathways for the conversion of GVL into fuels, fuel additive and chemical [10].
Membrane Reactors and Three-Phase Hydrogenation P a g e | 8
Polymer
Layer Membrane
Cell
conversion towards the desired product. In order for this to happen, we seek our reaction to be selective
towards
Figure 5. Matrimid chemical structure [12].
hydrogen, and that also removes selectively H2O. To accomplish such task, it was proposed the
development of a polymeric membrane based on the polymide polymer Matrimid (Figure 5). The
Matrimid solution was made and poured on a thin layer of polytetrafluoroethylene (PTFE) with pores of
a size of 0.05 micrometers. PTFE serves as the support for the polymer. The PTFE helps the membrane
be selective toward gas and liquid separation, which is one of the properties that we seek in our
membrane to have.
We were seeking to increase the temperature to a range of 120-150oC, since this will help us
increase the rate of reaction. Currently, the only temperature that it could be reached with PTFE layer
coated with catalyst for the hydrogenation reaction was between the range of 70-80oC because the liquid
phase started permeating trough the membrane as the temperature increase, in other words, as the vapor
pressure increased. Figure 6 shows the basic idea of how the membrane would look like inside the
membrane cell. For the three phase reaction of levulinic acid (LA), the liquid side will come in contact
with dense layer of the polymer, and the gas phase will be exposed to the porous side of the membrane.
Liquid Phase
Gas Phase
Porous
side
Polymer layer
PTFE layer
Dense
side
Figure 6. Matrimid membrane inside the membrane cell.
Membrane cell
Membrane Reactors and Three-Phase Hydrogenation P a g e | 9
Different procedure were executed when seeking to produce a more efficient membrane. One of
them, for example, was dissolving the Matrimid polymer in methylene chloride (DCM) solvent and
letting it dry overnight to form a dense layer (Figure 7). DCM solvent was replaced by GBL and since
GBL don’t have the same volatility as DCM, the membrane was put in different bath, which consisted of
water and ethanol. When the membrane was submerge in ethanol, the polymeric layer detached. The
membrane also tended to become wrinkled, and if the catalyst solution was poured, it tended to gain a
rough surface (Figure 8). The best fluid to harden the membrane was water. During the trials, I observed
that when the membrane was submerge in water without the catalyst solution, the desired structure was
obtained (Figure 10), the shiny dense layer. A problem notice when submerging it coated with catalyst
was that the dense layer was lost along with some of the catalyst which was dissolved in water (Figure
9). A possible solution to this problem was to make a mixture of the catalyst solution with the polymer
solution. Throughout the research one of my focused was developing and improving the performance of
such membranes. The only problem developing the desired membrane (with the shiny dense layer) was
the thickness of the polymeric layer which represents a problem when trying to supply hydrogen.
Figure 8. PTFE coated with Matrimid polymer
dissolved in GBL. Membrane soaked with
ethanol. Left membrane was coated with catalyst
solution and the right one only contains the
polymer solution.
Figure 10. PTFE coated with Matrimid polymer dissolved
in GBL. Catalyst solution was mixed with the polymer
solution. Membrane was soaked with water.
Figure 7. PTFE coated with Matrimid polymer
dissolved in DCM. Membrane was left to dry
overnight.
Figure 9. PTFE coated with Matrimid
polymer dissolved in GBL. Catalyst solution
was poured throughout the polymer solution.
Membrane was soaked with water.
Membrane Reactors and Three-Phase Hydrogenation P a g e | 10
The last type of membrane developed consisted of Matrimid dissolved in DCM and instead of
letting the solution dry overnight, it was spin coated into the PTFE (Figure 11). It is important to be able
to detect any degree of defect of the membrane. An imperfection in a membrane can cause a
nonselective convective flow pathways for gases or liquids through the membrane.
1.5 Research objectives summary
Traditional three-phase reactors often present mass transfer limitations, in other words hydrogen
delivery to the catalyst through the liquid phase. Hydrogen availability at the catalytic sites is often the
rate limiting step for hydrogenation reactions, which are industrially relevant in petrochemical and
petroleum refining and in emerging biorefining. Membrane reactors can lessen the mass transfer
limitations by directly and selectively supplying hydrogen to the catalytic sites on the membrane
surface. Therefore, the focus of this research was on developing and improving membranes for the
hydrogenation of LA.
Figure 11. PTFE coated with Matrimid
dissolved in DCM. Spin coater was used to dry
the solution on the surface.
Figure 12. Membrane Reactor Scheme [Image provided by John P. Stanford].
Membrane Reactors and Three-Phase Hydrogenation P a g e | 11
2. Research Methodology
2.1 Membrane Fabrication
Method 1. This type of membrane was made by crushing the pores of the PTFE. Catalyst solution
consisted of 0.2wt% RuCl3 dissolved in ethanol. This solution was poured on top of the membrane and
was then left to hydrogenate for two hours to ensure that the only substance on the surface was Ru.
Method 2. A polymeric solution consisting of 34g Matrimid, 74g THF, 74g GBL and 18g of butanol
was poured on top of the PTFE layer and a casting knife was used to ensure a smooth and even surface.
Then, it was submerge in water until the polymer harden. After the water treatment, a catalyst solution
consisting of 0.2wt% RuCl3 dissolved in ethanol was poured on top and the membrane was
hydrogenated for two hours.
Method 3. Two solutions of Matrimid dissolved in DCM were made. One consisted of 1wt% and the
other of 0.25wt%. They were poured on the PTFE layer and were spin coated at a 2500rpm for 60s. The
amount of 20uL, 50uL and 100uL were poured on the PTFE/Matrimid membrane. Each one of them was
hydrogenated for two hours.
Method 4. This type of membrane was made by preparing a 20wt% solution of Matrimid dissolved in
GBL mixed with a 0.5wt% RuCl3 catalyst solution dissolved in GBL. A casting knife was used to ensure
a smooth, tin and even surface. The membrane was submerge in water until the polymer harden. It was
then coated with 0.2wt% RuCl3 dissolved in ethanol and it was hydrogenated for two hours.
Method 5. This membrane consisted of PTFE coated with the catalyst solution which consisted of
0.2wt% RuCl3 dissolved in ethanol.
2.2 Calibration Samples
GBL calibration sample was prepared by mixing GBL with water. The following approximate
concentration were used: 0.10, 0.25, 0.5, 0.75, 1.0 and 1.5wt%. LA calibration samples were also
prepared by mixing LA with water using the following concentrations: 0.25, 0.5, 0.75, 1.0, 1.5 and 2.0wt
%.
Membrane Reactors and Three-Phase Hydrogenation P a g e | 12
2.3 Catalyst solution
The catalyst solution consisted of 0.2wt% RuCl3 dissolved in ethanol. The other catalyst solution
consisted of 0.5wt% RuCl3 dissolved in GBL.
2.4 GBL and water permeability
GBL and water permeability was tested in the membrane reactor to determine which fluid is
more convenient to use for the reaction. PTFE layer was placed in the membrane cell and was exposed
to the liquid for approximately 12 hours. Different temperature were tested, such as 60o, 70o and 80oC.
2.5 Solvent solutions
2.1g of Levulinic acid was dissolved in 70g of the following solvents: water, GBL and glycerol.